Sparx Systems dialog with customers

[roates]

I find the avenues for suggesting features for Enterprise Architect very limiting. At some risk of being flamed, I’m going to explain what mean.

To request a feature the [email protected] email address can be used but there is no way to monitor the progression of the suggestion. All you can do is sit and wait for X time to see if the feature ever gets implemented.

To provide a general comment the [email protected] email address can be used with the subject of EA general comments.

New features get added periodically and various bugs get fixed with each new release but there doesn't seem to be a dialog between the users and Sparx about what features are being looked at or what deficiencies are on the roadmap to be addressed.

I realise there is certainly an aspect of commercial confidentiality involved and that Sparx don't want to necessarily telegraph their every move but still, the way things are currently run seems pretty arrogant.

As a member of an EA team of a government department that has settled on using TOGAF 9.1 and ArchiMate 2.0 with minimal modification to support its EA efforts, I’m very disappointed with Sparx Enterprise Architect’s support for these two standards. The disappointment is particularly felt as I was one of the chief proponents for adopting Sparx Enterprise Architect as a modelling tool a number of years back.

TOGAF 9.1
The MDG Technology provided by Sparx at an additional cost, while perhaps operationally adequate, is a very bare bones offering. There are minimal if any example views provided. Not a single example pre-canned report and no guidance or roadmap around how to approach gap analysis as well as no exploitation of the included generic document templates in terms of the potential to auto-populate them with some data.

I downloaded it, emailed Sparx about the complete lack of included reports (with no useful result except confirmation that no reports are included), evaluated it from the point of view of how well some of our less experienced Sparx users in the team would be able to become productive with it and, chose not to pursue it further. A pity because we could certainly use a good TOGAF offering from Sparx – just not the current one they’re selling.

ArchiMate 2.0
The support for ArchiMate 2.0 is incomplete. Missing is:
- Some of the shapescripts (Representation, Business Event, Node ) need QA - the text is middle-vertical aligned so that when drawing nested elements the name of the containing element is obscured if it is one of the offending three. (Reported as a bug a long time ago to Sparx)
- Support for extension
- Support for implied relationships
- Support for nesting is not included in line with the ArchiMate 2.0 specification. The behaviour should be that when dropping an element onto another element a dialog should appear allowing the appropriate relationship to be selected from a set of allowable relationships. In EA this is not implemented and instead the parent element “owns” the element that was dropped onto it. (For an example of how it should ideally work, download Archi from http://archi.cetis.ac.uk/ and try out nested elements)

On the point of nesting, the Open Group ArchiMate® Tool Certification Conformance Requirements Version 2.0 January 2012 document states “A conforming product shall support nesting as an alternative representation of relationship types as described in Sections 7.1.1, 7.1.2, and 7.1.3 of the ArchiMate Specification. The conforming product shall clearly indicate which relationships are defined by each nesting instance, and, in updatable views, shall enable user control of relationships to be created, modified, or deleted.
A conforming product shall ensure that the graphical notation used for ArchiMate concepts and relationships remains unambiguously compliant with the ArchiMate Specification even after changes to the size, proportion, or color of modeling symbols.”

Although Sparx doesn’t claim that Enterprise Architect is ArchiMate® 2 Certified, they certainly don’t shy away from implying that the tool supports ArchiMate. Reading http://www.sparxsystems.com.au/products/mdg/tech/archimate/archimate.html gives an impression of support for the standard. ArchiMate is also included under the heading Major Supported Standards on the Sparx home page (left hand column towards bottom of page).

What I don’t understand is, why Sparx won’t go the extra mile? Enterprise Architect is a very good product. It has a lot of capabilities and I’m sure it has a very dedicated user base. Why then this approach of only doing some things 90%.

What do other users think?

Is Sparx engaging with it’s user base in the best way possible?  Can it do a better job? I hope so but will wait (a bit) and see.

[Geert Bellekens]

Hi roates,

You are no the first (and I'm sure you won't be the last) to complain about that.
There is indeed no public road-map, and no follow system for request features or reported bugs.
In the past Sparx has always been very silent when questioned about this.

I can't say anything about the support for Togaf and Archimate (because I don't use it), but in general my experience is that Sparx is rather willing to add any features needed to comply with the supported standard, but you do have to ask for it.

About the lack of reports, well l don't see that as a real problem.
I've never used any of the "standard" reports, and I think you almost always need to create your own.
Reporting is probably an area where you may want to invest in a consultant to create precisely that what you want. Given that the license price of EA is only 10% of that of IBM's rational suite, there should still be room in your budget

Geert

[qwerty]

I gave up hope some years ago.

q.

[Sunshine]

As an Enterprise Architect I too use ArchiMate. I have also found the support of ArchiMate lacking in some areas. I agree it would be good if ArchiMate 2.0 shape scripts were improved as I've had to cludge my way through the short falls. For instance before ArchiMate 2.0 was supported I had actually created my own MDG to support motivations and implementation extensions. It took up a few days of my time which could be better spent on other things. I'm sure to fix the problems roates mentioned above wouldn't take more than a few days of development effort so its not a heavy investment in time to polish off a good product and make it great.

I know Enterprise Architect is a UML based tool but its rather ironic don't you think it doesn't properly support an Enterprise Architecture Notation like ArchiMate. Come on Sparx make Enterprise Architect live up to its name.

As the Open Group are merging ArchiMate Notation into TOGAF I'm not using the TOGAF extension. I tried it and found it just too complicated for my team's to use.

At the risk of being given an inch and taking a mile I think a really cool feature would be to add a set of wizards and patterns that create the TOGAF lifecycle and under each phase provide example viewpoints. A lot of these are documented on the Open Group Website. This would make it easier for those learning ArchiMate with a set of example viewpoints. I appreciate this is a whole load more effort but it would be a real selling point and provide a good value proposition over other tools.

Well thats my 2 cents worth.  

[roates]

Thanks for the comments so far

Geert:

in general my experience is that Sparx is rather willing to add any features needed to comply with the supported standard, but you do have to ask for it.

I'm more coming from the perspective that Sparx have chosen to implement ArchiMate 2.0 but haven't finished the job.
As for requesting features... I previously pointed out the shapescript issues and posted in the forum some time back about the lack of extensibility. On the shapescript issue I did get a reply to the effect that it would be fixed - but it hasn't... On the extensibility issue got nothin'.

The point on the lack of reports for the TOGAF MDG Tech is basically that I think it would round out the offering. At $145 per license what you're getting is essentially a profile and some locked down graphics.
Download it and check it out.

As for the perennial argument that Sparx is cheap and so we should forgive its inadequacies... I don't think that holds much water anymore. I would prefer Sparx to do some rigorous QA on what is currently in the product rather than getting an email soon about

Enterprise Architect 9.5 beta 1 is out!

followed 3 weeks later with

Enterprise Architect 9.5 is released!

, knowing pretty much full well that the existing holes will probably still be there.

Don't get me wrong, I like Sparx Enterprise Architect. I've championed it at 3 workplaces in the last decade because it's a pretty good product and it's Australian  . I want to see it succeed.

Qwerty:
I think we’re on the same page. Have a look at Archi, only to get a glimpse of the additional functionality that Sparx could implement if they went the extra mile. With Archi, doing a technology layer diagram is a breeze because of its nesting ability. The diagram just looks so much better and as you know, often the simplicity of the diagram is a factor when you’re communicating to certain stakeholders. Trying to do the same thing (nesting) in Sparx doesn’t work for nesting because of the current shapescript issues but even if the shapescripts were fixed, the associations between elements would first have to be manually created and then the elements stacked and then the connectors hidden on the diagram so as they then won’t show. A lot of mucking around/working around.

I agree with your suggestions about wizards and views related to the TOGAF ADM. Even if Sparx doesn’t do that stuff, if they’d fix the support for Archimate 2.0 then it would be possible for users to build some additional functionality on top by way of MDG tech etc but as it is, to invest on building stuff on top of an incomplete implementation just doesn’t make a lot of sense. There is a great deal of scope for Sparx to beef up the product in the area of supporting practising EAs though, I agree. (Imagine if they began a roadmap to build in the kind of inline live editing of document arifacts to repository ala Mega (http://www.mega.com/en/c/solution/p/ea/p2/enterprise-architecture).

Anyway, thanks for the responses. I hope we get some more discussion and ultimately a response from Sparx that hopefully isn’t defensive but actually welcoming of considered feedback.

[Thelonius]

Good conversation. We all love Sparx, or loved it lots in the beginning and have gradually been falling out of love with it in recent years.

Based on Sparx's corporate behaviour over the last six years, I have concluded that while there are a number of really great people in Sparx who care and are doing great work, the people who own the company are out of touch and don't seem to be aware (or don't care) what's happening in the global market place or what their customers would happily pay lots of money for.

Sparx could be so much better than it is now. I just don't see any evidence that they have the strategy or the intent to improve. Or that they're even aware of the opportunity.

If I were wrong, and Sparx suddenly pulled a rabbit of their hat - I'd be elated. But I've lost hope and moved on. I have met at least 10 corporate customers of Sparx in the last three years who have done the same. Met one last week, in fact. A big one.

Suddenly the image of Gerry Harvey just popped into my mind for some reason.

[RoyC]

Yes, this is a good conversation. Perhaps you could officially direct specific questions to Sparx Support (include a link to this thread) so that the appropriate people are also aware of the points you are making. As Geert points out, we like to be responsive, but one or two comments here suggest that maybe Support have not received a question to be responsive to!

[roates]

Hi Royc,

OK, that's fair enough.
I've found 13 automated responses to bug/feature requests that I submitted since 2008 but I’ll check which of the items I've highlighted in the top post haven't actually been put into the bug system previously.

I'll enter each item as a bug, not a feature request as I'm pretty sure that feature requests receive nothing but an automated response. (correct me if I'm wrong...)

The bugs I'll enter are for now are:

1. Shapescript title placement for ArchiMate 2.0 elements (Representation, Business Event, Node) need to be verticallay aligned towards the top of the element.

2. ArchiMate 2.0 cannot be extended. From document, Open Group, ArchiMate Tool Certification – Conformance Requirements Version 2.0 January 2012:

2.5 Language Extension Mechanisms
A conforming product shall permit users to add attributes to ArchiMate concepts and relationships as described in Section 9.1 of the ArchiMate Specification.
A conforming product shall permit users to define specializations of ArchiMate concepts as described in Section 9.2 of the ArchiMate Specification.

3. Nesting behaviour is incorrect. From document, Open Group, ArchiMate Tool Certification – Conformance Requirements Version 2.0 January 2012:

2.3 Language Notation
A conforming product shall implement the vocabulary, notation, syntax, and semantics of the visual modeling language described in Chapters 3-7, Chapters 11-12, and Appendix A of the ArchiMate Specification.
A conforming product shall support nesting as an alternative representation of relationship types as described in Sections 7.1.1, 7.1.2, and 7.1.3 of the ArchiMate Specification. The conforming product shall clearly indicate which relationships are defined by each nesting instance, and, in updatable views, shall enable user control of relationships to be created, modified, or deleted.

4. Connectors
Drawing a connector between one ArchiMate 2.0 element and another ArchiMate 2.0 element results in the Line Style of the connector defaulting to Orthogonal Rounded, irrespective of the Tools | Options | …. Um… forget it. I’m pretty sure a global connector line style could be set via Tools | Options but hunting for it now it appears to have been removed….

4. Derived Relationships not functional. From the document, Open Group Standard ArchiMate® 2.0 Specification:

7.5 Derived Relationships
… Part of the language definition is an abstraction rule that states that two relationships that join at an intermediate element can be combined and replaced by the weaker of the two.
If two structural relationships r:R and s:S are permitted between elements a, b, and c such that r(a,b) and s(b,c), then a structural relationship t:T is also permitted, with t(a,c) and type T being the weakest of R and S.
For the application of this rule, it is assumed that the assignment relationship has a direction (as indicated by the role names in Figure 2, Figure 3, Figure 9, Figure 26, Figure 34, and Figure 44).
Transitively applying this property allows us to replace a “chain” of structural relationships (with intermediate model elements) by the weakest structural relationship in the chain. For a more formal description and derivation of this rule we refer to [13].
With this rule, it is possible to determine the “indirect” relationships that exist between model elements without a direct relationship, which may be useful for, among other things, impact analysis.

Thanks for replying on the thread.

[skiwi]

I agree with and support the tenor of this thread.

I too have previously expressed a wish for a more transparent and accessible tracking of bugs, including publishing resolved bug identifiers in the release notes.

I feel that in a number of areas (and recently I saw in these threads the phrase "sales ware") a new feature is implemented in a bare bones fashion, lacking documentation (Strategy Maps?) and then little is done to enhance and grow it.

Several come to mind
* org charts
* diagramming, esp white board and hand drawn charts and the symbols provided, text formatting, layout, and alignment, usability; (see other threads)
* rtf documentation (well perhaps not bare bones, but definitely in need of continued development and support)
* a forum where I could search back general discussions more than a year without timing out, get more than 100 results, ... (not EA per se, but symptomatic)

[Wyatt]

A workaround for forum search from http://www.sparxsystems.com/cgi-bin/yabb/YaBB.cgi?num=1330884294/8#8:

http://groups.google.com/forum/#!forum/sparx-enterprise-architect-general
http://groups.google.com/forum/#!forum/sparx-enterprise-architect-uml-process
http://groups.google.com/forum/#!forum/sparx-enterprise-architect-bugs-and-issues

[skiwi]

A workaround for forum search from http://www.sparxsystems.com/cgi-bin/yabb/YaBB.cgi?num=1330884294/8#8:

http://groups.google.com/forum/#!forum/sparx-enterprise-architect-general
http://groups.google.com/forum/#!forum/sparx-enterprise-architect-uml-process
http://groups.google.com/forum/#!forum/sparx-enterprise-architect-bugs-and-issues

Nice, perhaps all forums could be moved ... !
(I note the bugs and issues is by invite only) now fixed
Be good to have suggestions and requests as well

[d.greenyer-sf]

I have to agree with this post, I really like the EA product but Sparx seem to have lost their way and their market advantage.

I believe we need a focus away from Software Architecture and more focus on Enterprise Architecture.

A move to make EA more web enabled.

Better visualisations.

Better MDG technologies.

We need a more modern EA that is back in the Gartner magic quadrant and can compete with Bizzdesign, Orbus and Troux and the rest.